A c-theorem for the effective central charge in the R=1 replica limit, and applications to systems with measurement-induced randomness

Abstract: We present a general theorem demonstrating non-perturbatively the decrease of the "effective central charge" $c_{\text{eff}}=(d c/dR)|{R=1}$ under renormalization group (RG) flow in the $R\rightarrow1$ replica limit of a $R$-copy $2D$ conformal field theory (CFT) action $S{}$ perturbed by a replica interaction of the form $$-\mathbb{S}=-\sum_{a=1}^{R}S_{}^{(a)}+\Delta\int d^2 x \sum_{\substack{a,b=1\ a\neq b}}^{R}\varphi^{(a)}(x)\varphi^{(b)}(x).$$ Here $\varphi$ is a scaling field belonging to the CFT with action $S_$ and the coupling $\Delta$ is relevant in the RG sense. We show that the infrared value of $c_{\text{eff}}$ is always $\textit{less}$ than the central charge $c$ of the unperturbed CFT $S_{}$. We refer to this result as the "$c$-effective theorem". As an application of this theorem, we consider replica field theories in the limit of $R \to 1$ replicas of the form above, shown by Nahum and Jacobsen [arXiv:2504.01264] to describe $2D$ classical monitored systems, where measurements introduce a form of quenched randomness via Bayes' theorem. Lastly, we discuss a possible relationship of our theorem with the effective central charge $c_{\text{eff}}^{(R\rightarrow0)}=(dc/dR)|_{R=0}$ for the above replica action in the different $R\rightarrow0$ replica limit, which is of relevance to systems with generic uncorrelated impurity-type quenched disorder, as opposed to measurements.